Accordingly, this program uses set of prime numbers for parallel generating of Normal random sequence. Reference ( ) suggest a method of choosing the additive constant c for parallel generation.
MONTE CARLO METHOD EXAMPLES FULL
In addition, for the example to utilize the full power of the multiple DSP accelerators, the merging together of multiple independent random sequences into a single random sequence and still preserve the good “randomness” attributes is an issue. Thus the constants a and M of formula were chosen to ensure a long pseudo random sequence.
MONTE CARLO METHOD EXAMPLES HOW TO
The issue of how to choose the constants a, M and c and the initial seed x0 to achive good “randomness” features is discussed in many publication, see ( ) for example. Measure the Entropy - lack of order or predictability For a finite sequence X, the entropy can be defined as the following.
No predictability (After reaching the length of the sequence starts to repeat itself, so the longer the sequence, the better) Have the correct distribution (lim (1/n * S (X(n)) ~ N(0,1)
Randomness, sequence length and parallel computation ¶ Multiple tests were suggested in the literature for the “randomness” of a Normal Random sequence:
0 kommentar(er)
